The adenosine 2A receptor agonist CGS-21680 (2-p-[2-carboxyethyl] phenethyl-amino-5′-N-ethylcarboxamido-adenosine) was obtained from Tocris bioscience (Cat #1063). Polyclonal goat antibodies against tPA (Cat #387), uPA (Cat #398) and PAI-1 (Cat 395G) were obtained from American Diagnostica inc.; monoclonal mouse antibody against annexin A2 from BD Transduction Laboratories (Cat #610069); rabbit antibody against ß-actin from Sigma Chemical Co. (Cat #A2066); Alexa fluor^® 488 goat anti-mouse (Cat #A1101) antibody from Molecular Probes Ltd. All other materials were the highest quality that could be obtained.

Human dermal microvascular endothelial cells (HDMVEC) from neonate foreskin (Lonza, Cat #CC-2505) were cultured up to 80–90% confluence in supplemented EGM-2MV medium (Lonza, Cat #CC-3202) at 37°C and 5% CO₂ in a humidified atmosphere. Cells used in all experiments were between passages 5 and 7. The night prior the experiments, medium was replaced to 0.1% Bovine Serum Albumin (BSA, ELISA grade, Sigma Chemical Co., Cat #A7030) supplemented EBM-2 medium (Lonza, Cat #CC-3156). All experiments were carried out in 0.1% BSA basal medium.

HDMVEC, cultured in 24 well plates, were incubated in the absence (control) or presence of increasing concentrations (10⁻⁷–10^–5 M) of the selective A_2A adenosine receptor agonist CGS-21680 in 0.1% BSA basal medium for 24 h at 37 °C and 5% CO₂. Supernatants were collected and frozen at -80 °C until ELISA determination. Cell viability was determined by cellular reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, 0.2 mg/ml, Sigma Chemical Co. Cat #M2003) to formazan, solubilization in DMSO and absorbance measurement at 490 nm. Concentration of total tPA, uPA and PAI-1 antigens present in conditioned media were determined using commercially available assays: IMUBIND^® tPA (Cat #860), uPA (Cat #894) or PAI-1 (Cat #822) ELISA (American Diagnostica Inc.). Sensitivity of the kit for tPA is 0.2 ng/ml, for uPA is 0.02 ng/ml and for PAI-1 is 1 ng/ml. Results were normalized to 10⁵ viable cells per well.

Total RNA (1 µg) isolated from cultured cells using Trizol reagent (Invitrogen Cat# 15596018) was transcribed into cDNA (RT) using the RNA PCR Core Kit (Applied Biosystems™ N8080143). Aliquots of RT were subjected to real-time PCR using the Mx3005P system (Stratagene, La Jolla, CA), SYBR-Green Brilliant Master Mix (Stratagene Cat #600548) and specific primers for t-PA (Forward 5′-CCC​AGA​TCG​AGA​CTC​AAA​GC-3′ Reverse 5′-TGG​GGT​TCT​GTG​CTG​TGT​AA-3′, NCBI Reference Sequences: NM_000921 and NM_033011), u-PA (Forward 5′-ATT​CAC​CAC​CAT​CGA​GAA​CC-3′ Reverse 5′-TCC​ACC​TCA​AAC​TTC​ATC​TCC-3′, NCBI Reference Sequence: NM_002658), PAI-1 (Forward 5′-CTG​GTT​CTG​CCC​AAG​TTC​TC–3′ Reverse 5′-GAC​TGT​TCC​TGT​GGG​GTT​GT-3′, NCBI Reference Sequence: NM_000602), and GAPDH (Forward 5′-AAC​ATC​ATC​CCT​GCC​TCT​AC–3′ Reverse 5′-CCC​TGT​TGC​TGT​AGC​CAA​AT-3′); Annealing Temperature: 60°C. All values were normalized to GAPDH as described (Che et al., 2007). All primers were designed in a way that the PCR products spanned an intron.

HDMVEC were grown on chamber slides (Nalge Nunc International, Cat #177402). After treatment, medium was discarded, and cells were fixed in ice cold 70% ethanol and blocked with 2% BSA in 0.01% (vol/vol) Tween 20 PBS - (PBST). Slides were incubated with primary antibodies against proteins of interest and subsequently incubated with Alexa Fluor 488 secondary antibodies, nuclei were counterstained with Hoerschst 33342 (Molecular Probes Inc. Cat #H21492) and slides were mounted in fluorescence mounting medium (DakoCytomaton, Glostrup, Denmark). Negative controls were run in parallel without the primary antibody. Fluorescence was detected with a Nikon Eclipse e800 microscope coupled with an automated imager Nikon ACT-1 (Nikon Instruments Inc. Melville, NY) and fluorescence intensity was blindly quantified with MetaMorph software (Molecular Devices Corporation, Downington, PA) by an independent observer.

HDMVEC were seeded on poly-l-lysine (100 μg/ml) coated 24 mm diameter glass coverslips in six-well plates. After treatment, calcium-dependent translocation of annexin A2 to the membrane (Monastyrskaya et al., 2007) was induced by addition of 10⁻⁶ M ionophore A23187 (Sigma Chemical Co., Cat #C7522) for 5min. Then, medium was discarded and cells were fixed in ice cold 70% ethanol, blocked with 2% BSA in 0.01% (vol/vol) Tween 20 PBS - (PBST), incubated with primary antibodies against annexin A2, subsequently incubated with Alexa Fluor 488 goat anti-mouse antibody and nuclei were counterstained with Hoerschst 33342. Confocal images were obtained on a LEICA TCS SP2 (DM-IRB) laser-scanning microscope. A z-series of images were acquired at 2 µm steps to produce individual z-stacks. The resulting images were overlaid and analyzed using the Leica software, v2.61 (Leica Geosystems AG, Heerbrugg, Suiza).

HMVEC, cultured in 6 well plates, were incubated in the absence (control) or presence of the selective A_2A adenosine receptor agonist CGS-21680 in 0.1% BSA basal medium for 24 h at 37°C and 5% CO₂. Supernatants were removed and cells were rinsed in ice-cold phosphate-buffered saline (PBS), pH 7.4 and lyzed at 4°C in lysis buffer (50 mM Tris pH 7.4, 150 mM NaCl, 0.1% SDS, 1% Nonidet P-40 and 0.5% sodium deoxycolate) plus protease inhibitors (leupeptin, aprotinin and PMSF). Following centrifugation (10000g, 10 min) equal amounts of protein (5 or 25 µg/lane) were separated by 10% SDS-PAGE under reducing conditions and electrophoretically transferred onto poly-(vinylidene difluoride) (PVDF) membranes, blocked with 3% nonfat milk in 0.1% (vol/vol) Tween 20 phosphate buffered saline (PBST) and then incubated with specific antibodies against the protein of interest. After extensive washes, blots were incubated with a horseradish peroxidase-conjugated secondary antibody and the immunoreactive bands were visualized by enhanced chemiluminescence (Amersham Biosciences, Cat #RPN2232) using the AutoChemi image analyzer and Labworks 4.6 software (UVP, Inc. Upland, CA).

After treatment, HDMVEC cultured in T25 flasks were lyzed with 1 ml of lysis buffer (50 mM Tris pH 7.4, 150 mM NaCl, 0.1% SDS, 1% Nonidet P-40 and 0.5% sodium deoxycolate) containing protease inhibitors (leupeptin, aprotinin and PMSF). Whole cell lysates were precleared for 30 min at 4°C with 0.25 µg of appropriate control (normal) IgG, corresponding to the host species of the primary antibody (Mouse IgG, Sigma Chemical Co., Cat #I5381), and 20 µL of protein A/G-agarose (Santa Cruz Biotechnology, Inc. Cat #sc-2003). The precleared cell lysates were subsequently incubated with 2 µg of the monoclonal antibody against annexin A2 for 2 h at 4°C before the addition of 20 µL of protein A/G-agarose, and the incubation was continued overnight at 4°C with constant rotation. Next, the agarose beads were pelleted by centrifugation at 2500 rpm (1,000g) for 5 min at 4°C. After extensive washes with ice-cold lysis buffer, beads were resuspended in sample loading buffer, boiled for 4–5 min, and subject to Western blotting with specific antibody against tPA.

Plasminogen activation was determined as the ability of generated plasmin to degrade a specific fluorogenic substrate by membrane bound exogenous tPA (Diaz et al., 2004). HDMVEC (10⁴ cells/well) were cultured in 96 well microplate, black, clear bottom (Corning Incorporated, Cat #CLS3603). After treatment, supernatants were removed, and cells were washed twice with PBS and incubated with 10 nM of single-chain recombinant tPA (American Diagnostica inc. Cat #173) in PBS-BSA 2% for 30 min. After washing three times with PBS to remove unbound tPA, 100 nM Glu-plasminogen (Calbiochem, manufactured by EMD Biosciencies, Inc.; Cat #528180) and 50 µM of the fluorogenic substrate D-Ala-Leu-Lys-7-amido-4-methylcoumarin (Sigma Chemical Co., Cat #A8171) in buffer 0.05 M Tris-HCl, 0.1 M NaCl, 0.01% Tween 20, pH 7,4 were added. Plasmin formation was monitored at 360-nm excitation/460-nm emission setting in a spectrofluorimeter plate reader (Perkin-Elmer WALLAC Victor² 1420, Perkin-Elmer life sciencies, Turku, Finland) at 5- or 30-min intervals for 6 h at 37°C. The basal fluorescence obtained at 0 min was subtracted from each time point. Additionally, we performed a control chemical reaction without cells to monitor the kinetics of the reaction. To determine the role of plasminogen binding to HDMVEC, these assays were also performed including 6-aminocaproic acid (6-ACA, Sigma Chemical Co., Cat #A2504; 1 mM) in the final reaction mixture as antagonist of this binding.

200 µL of HDMVECs cell suspension (5 × 10⁴ cells/ml from passage 4) in EGM-2MV medium were seeded on 50 µL of polymerized (37°C for 30 min) Matrigel^® (BD Biosciences, Cat #354234) in a 96 well plate. After 18 h incubation (37°C and 5% CO₂), Fluorophore/calcein AM (Molecular Probes, Inc. Cat #C1430) was added to stain cells. Image acquisition of tubular network was achieved by fluorescence microscopy (Nikon Eclipse e800 microscope coupled with an automated imager Nikon ACT-1, Nikon Instruments Inc. Melville, NY) and Sigma Scan Pro software (SPSS, Chicago, IL) was used to determine tube surface area (Perez-Aso et al., 2014; Vicente et al., 2016).

Differences between groups in the in vitro studies were analyzed by means of one-way analysis of variance (ANOVA) by Dunnet’s multiple comparison test performed by GraphPad Prism 4 software (GraphPad Software, Inc. San Diego, CA).

The expression of the key components of the plasminogen activation system, uPA and its receptor (uPAR), as well as its inhibitor PAI-1, is induced early during re-epithelialization in the migrating epithelial sheet in incisional wounds in mice (Sulniute et al., 2016). In contrast, the vascular endothelium is considered to be the major site of synthesis and release of tPA (Emeis et al., 1997). Since we have previously shown that adenosine A_2A receptor activation promoted wound revascularization by promoting both angiogenesis and vasculogenesis (Montesinos and Valls, 2010), we investigated its effect on human dermal microvascular endothelial cells (HDMVEC).

HDMVEC secreted both types of plasminogen activators, tPA and uPA, in lesser amounts than their inhibitor PAI-1 (Figures 1A–C). The addition of increasing amounts of the selective A_2A receptor agonist did not produce any appreciable change in the secretion of tPA and uPA, but significantly diminished the levels of PAI-1 present in the conditioned media of HDMVEC (79.3 ± 12.1 ng/ml in 10⁻⁶ M CGS-21680 treated vs. 171.2 ± 9.3 ng/ml in untreated HDVECs, ***p < 0.001, n = 4, Figure 1C). Real time RT-PCR determination of the mRNA levels for tPA, uPA and PAI-1 showed that A_2A adenosine receptor activation did not affect tPA and uPA mRNA expression in HDMVECs while decreasing PAI-1 message levels (39% reduction from control by 10^–6 M CGS-21680, p = 0.050, n = 3, Figures 1D–F). Therefore, the effect of adenosine A_2A receptor activation on the level of uPA, tPA and PAI-1 expression correlated with the effect on released proteins to the condition media.

Since we had observed that CGS-21680 treatment did not affect the expression and the release of tPA by HDMVEC, we decided to investigate the production of annexin A2, a protein that acts as an endothelial cell membrane co-receptor of plasminogen and tPA (Hajjar et al., 1994). Annexin A2 is an abundant protein (∼36 KD) in HDMVECs with mainly perinuclear localization. The activation of A_2A adenosine receptors by CGS-21680 significantly increased the overall levels of this protein, shifting its localization toward the cellular membrane (Figures 2A,C). These findings were further confirmed by confocal laser scanning microscopy (Figure 2D).

In order to determine whether there is any physical interaction between endogenous tPA and annexin A2, we immunoprecipitated annexin A2 and observed that the amount of tPA (∼60 KD) that co-precipitated with annexin A2 increased markedly in the CGS21680-treated cells (Figure 2B).

We determined the functionality of the increased annexin A2 production by an indirect test of plasminogen activation, through the degradation of a specific substrate for the newly generated plasmin. As a control, we performed the assay in the absence of cells and confirmed the selectivity of the substrate, which was only degraded after activation of plasminogen by recombinant tPA (Figure 3A). We observed that the low amounts of plasminogen activators, tPA and uPA, endogenously produced by both untreated and adenosine A_2A agonist treated HDMVEC were not sufficient to activate plasminogen to plasmin (Figure 3B). To detect exclusively membrane-bound tPA-associated plasminogen activation it was necessary to incubate exogenous recombinant tPA for 30 min and then washed away before the addition of substrate and Glu-plasminogen. The lag time observed in plasmin generation by HDMVEC-surface bound tPA, about 20–30 min, is similar to that described for cerebral microvascular endothelial cells (Semov et al., 2005), but much longer than for pancreatic or breast cancer cells (Diaz et al., 2004; Sharma et al., 2010).

CGS-21680 pre-treatment produced an increase in membrane bound tPA-associated plasmin generation by HDMVEC (Figures 3B,C). Both preincubation with a specific antibody against annexin A2 (4 μg/ml) before binding of recombinant tPA, or the presence of the plasmin inhibitor 6-ACA (1 mM) in the final reaction, markedly inhibited plasmin generation and completely abolished the CGS-21680-mediated increase in cell associated-plasmin generation (Figure 3C).

Once we established that adenosine A_2A receptor activation increased annexin A2 expression resulting in an increase in cell-bound tPA and thereby accelerated plasminogen activation, we evaluated the contribution of this system in another functional assay, the endothelial cell formation of a tubular network on a three-dimensional matrix. In a previous study, we demonstrated that A_2A receptor activation stimulates endothelial network formation in Matrigel and established the concentration dependent effect for the agonist CGS-21680 in this model (Desai et al., 2005). For the present study we selected the concentration of 1 μM, which had shown a maximal effect that could be abrogated by a selective A_2A antagonist (Desai et al., 2005). HDMVEC network formation was significantly inhibited by the plasmin inhibitor 6-ACA (1 mM) and a murine monoclonal antibody against annexin A2, but not a control murine antibody (Figure 4). In contrast, antibodies against tPA, uPA and PAI-1 did not affect basal network formation. Interestingly, adenosine A_2A receptor-mediated increases in network formation were blocked by 6-ACA and antibodies against PAI-1 and tPA but not uPA or annexin A2 (Figure 4).